# What is Tensormorph?

Tensormorph is an IDE for opening, navigating, inspecting, comparing, debugging, and versioning model checkpoints - from a few million parameters to trillion-parameter Mixture-of-Experts models - without needing to fully load a model into RAM or VRAM to work with it. If a debugger, a profiler, and an IDE had a shared tool for weight space instead of source code, this is roughly it.

## Who it's for

Anyone whose job involves looking closely at what's actually inside a checkpoint: diffing two fine-tunes to see what changed, diagnosing why a quantization pass hurt quality, tracking down where a NaN first appeared during training, or safely editing weights without corrupting a multi-hundred-gigabyte file by hand.

## How it's organized

Everything you inspect lives in one shared **address space** - model, architecture, layer, module, tensor, tile, row/column, down to a single scalar (see [Model, Architecture, Layer and Module](/docs/core-concepts/model-architecture-layer-module)) - and every view is a different way of looking at the same address space, kept in sync:

-   **3D views** for spatial structure: Architecture, Layer Stack, Tensor Volume, MoE Expert Galaxy, Parameter Map, Diff, Morph, Runtime Flow.
-   **2D views** for precise per-element reading: the [Matrix and Heatmap Editor](/docs/editors/matrix-and-heatmap-editor).
-   **1D views** for curves and distributions: the [Signal and Distribution Editor](/docs/editors/signal-and-distribution-editor).
-   The **Scalar Inspector**, for a single number in full detail.

Named **workspace presets** (Architecture, Compare, Morph, MoE, Runtime, Performance, and others) switch this whole layout at once for a given task, rather than requiring you to manually arrange panels each time.

## What it isn't

Tensormorph doesn't bundle its own version control. **Lob**, a separate tensor-native VCS, is what actually stores weight history; Tensormorph anchors to it (see [Merge Conflicts](/docs/compare-and-morph/merge-conflicts) for what that means in practice) rather than reimplementing it. Tensormorph is also not primarily a training tool - it reads checkpoints and runtime traces, rather than running training loops itself.

## Where to go next

Continue to [Install the Desktop Application](/docs/get-started/install-desktop-application) to get set up, or jump straight to [Open Your First Model](/docs/get-started/open-your-first-model) if Tensormorph is already installed.

## Related resources

-   [Model, Architecture, Layer and Module](/docs/core-concepts/model-architecture-layer-module) - the address space every view shares.
-   [Glossary](/docs/reference/glossary) - short definitions for every term introduced across this documentation.